Get our free extension to see links to code for papers anywhere online!

Chrome logo Add to Chrome

Firefox logo Add to Firefox
Picture for Csaba Szepesvári

Csaba Szepesvári

Optimization Issues in KL-Constrained Approximate Policy Iteration


Feb 11, 2021
Nevena Lazić, Botao Hao, Yasin Abbasi-Yadkori, Dale Schuurmans, Csaba Szepesvári

Figure 1 for Optimization Issues in KL-Constrained Approximate Policy Iteration
Figure 2 for Optimization Issues in KL-Constrained Approximate Policy Iteration
Figure 3 for Optimization Issues in KL-Constrained Approximate Policy Iteration
Figure 4 for Optimization Issues in KL-Constrained Approximate Policy Iteration

  Access Paper or Ask Questions

Bootstrapping Statistical Inference for Off-Policy Evaluation


Feb 09, 2021
Botao Hao, Xiang Ji, Yaqi Duan, Hao Lu, Csaba Szepesvári, Mengdi Wang

Figure 1 for Bootstrapping Statistical Inference for Off-Policy Evaluation
Figure 2 for Bootstrapping Statistical Inference for Off-Policy Evaluation
Figure 3 for Bootstrapping Statistical Inference for Off-Policy Evaluation
Figure 4 for Bootstrapping Statistical Inference for Off-Policy Evaluation

  Access Paper or Ask Questions

On Query-efficient Planning in MDPs under Linear Realizability of the Optimal State-value Function


Feb 04, 2021
Gellert Weisz, Philip Amortila, Barnabás Janzer, Yasin Abbasi-Yadkori, Nan Jiang, Csaba Szepesvári


  Access Paper or Ask Questions

Asymptotically Optimal Information-Directed Sampling


Nov 11, 2020
Johannes Kirschner, Tor Lattimore, Claire Vernade, Csaba Szepesvári


  Access Paper or Ask Questions

Sparse Feature Selection Makes Batch Reinforcement Learning More Sample Efficient


Nov 08, 2020
Botao Hao, Yaqi Duan, Tor Lattimore, Csaba Szepesvári, Mengdi Wang


  Access Paper or Ask Questions

Online Sparse Reinforcement Learning


Nov 08, 2020
Botao Hao, Tor Lattimore, Csaba Szepesvári, Mengdi Wang

Figure 1 for Online Sparse Reinforcement Learning

  Access Paper or Ask Questions

On Optimality of Meta-Learning in Fixed-Design Regression with Weighted Biased Regularization


Oct 31, 2020
Mikhail Konobeev, Ilja Kuzborskij, Csaba Szepesvári

Figure 1 for On Optimality of Meta-Learning in Fixed-Design Regression with Weighted Biased Regularization
Figure 2 for On Optimality of Meta-Learning in Fixed-Design Regression with Weighted Biased Regularization
Figure 3 for On Optimality of Meta-Learning in Fixed-Design Regression with Weighted Biased Regularization
Figure 4 for On Optimality of Meta-Learning in Fixed-Design Regression with Weighted Biased Regularization

  Access Paper or Ask Questions

Online Algorithm for Unsupervised Sequential Selection with Contextual Information


Oct 23, 2020
Arun Verma, Manjesh K. Hanawal, Csaba Szepesvári, Venkatesh Saligrama

Figure 1 for Online Algorithm for Unsupervised Sequential Selection with Contextual Information
Figure 2 for Online Algorithm for Unsupervised Sequential Selection with Contextual Information
Figure 3 for Online Algorithm for Unsupervised Sequential Selection with Contextual Information
Figure 4 for Online Algorithm for Unsupervised Sequential Selection with Contextual Information
* Accepted to NeurIPS 2020 
  Access Paper or Ask Questions

CoinDICE: Off-Policy Confidence Interval Estimation


Oct 22, 2020
Bo Dai, Ofir Nachum, Yinlam Chow, Lihong Li, Csaba Szepesvári, Dale Schuurmans

Figure 1 for CoinDICE: Off-Policy Confidence Interval Estimation
Figure 2 for CoinDICE: Off-Policy Confidence Interval Estimation
Figure 3 for CoinDICE: Off-Policy Confidence Interval Estimation
* To appear at NeurIPS 2020 as spotlight 
  Access Paper or Ask Questions

Exponential Lower Bounds for Planning in MDPs With Linearly-Realizable Optimal Action-Value Functions


Oct 03, 2020
Gellert Weisz, Philip Amortila, Csaba Szepesvári

Figure 1 for Exponential Lower Bounds for Planning in MDPs With Linearly-Realizable Optimal Action-Value Functions
Figure 2 for Exponential Lower Bounds for Planning in MDPs With Linearly-Realizable Optimal Action-Value Functions

  Access Paper or Ask Questions

Tighter risk certificates for neural networks


Aug 12, 2020
María Pérez-Ortiz, Omar Rivasplata, John Shawe-Taylor, Csaba Szepesvári

Figure 1 for Tighter risk certificates for neural networks
Figure 2 for Tighter risk certificates for neural networks
Figure 3 for Tighter risk certificates for neural networks
Figure 4 for Tighter risk certificates for neural networks
* Preprint under review 
  Access Paper or Ask Questions

Efficient Planning in Large MDPs with Weak Linear Function Approximation


Jul 13, 2020
Roshan Shariff, Csaba Szepesvári

* 12 pages and appendix (10 pages). Submitted to the 34th Conference on Neural Information Processing Systems (NeurIPS 2020), Vancouver, Canada 
  Access Paper or Ask Questions

Confident Off-Policy Evaluation and Selection through Self-Normalized Importance Weighting


Jun 18, 2020
Ilja Kuzborskij, Claire Vernade, András György, Csaba Szepesvári

Figure 1 for Confident Off-Policy Evaluation and Selection through Self-Normalized Importance Weighting
Figure 2 for Confident Off-Policy Evaluation and Selection through Self-Normalized Importance Weighting
Figure 3 for Confident Off-Policy Evaluation and Selection through Self-Normalized Importance Weighting
Figure 4 for Confident Off-Policy Evaluation and Selection through Self-Normalized Importance Weighting

  Access Paper or Ask Questions

Efron-Stein PAC-Bayesian Inequalities


Sep 04, 2019
Ilja Kuzborskij, Csaba Szepesvári


  Access Paper or Ask Questions

Detecting Overfitting via Adversarial Examples


Mar 06, 2019
Roman Werpachowski, András György, Csaba Szepesvári

Figure 1 for Detecting Overfitting via Adversarial Examples
Figure 2 for Detecting Overfitting via Adversarial Examples
Figure 3 for Detecting Overfitting via Adversarial Examples
Figure 4 for Detecting Overfitting via Adversarial Examples
* 25 pages 
  Access Paper or Ask Questions

Distribution-Dependent Analysis of Gibbs-ERM Principle


Feb 05, 2019
Ilja Kuzborskij, Nicolò Cesa-Bianchi, Csaba Szepesvári


  Access Paper or Ask Questions

Online Algorithm for Unsupervised Sensor Selection


Jan 15, 2019
Arun Verma, Manjesh K. Hanawal, Csaba Szepesvári, Venkatesh Saligrama

Figure 1 for Online Algorithm for Unsupervised Sensor Selection
Figure 2 for Online Algorithm for Unsupervised Sensor Selection
Figure 3 for Online Algorithm for Unsupervised Sensor Selection
Figure 4 for Online Algorithm for Unsupervised Sensor Selection
* Accepted at AIStats 2019 
  Access Paper or Ask Questions

LeapsAndBounds: A Method for Approximately Optimal Algorithm Configuration


Jul 02, 2018
Gellért Weisz, András György, Csaba Szepesvári

Figure 1 for LeapsAndBounds: A Method for Approximately Optimal Algorithm Configuration
Figure 2 for LeapsAndBounds: A Method for Approximately Optimal Algorithm Configuration
Figure 3 for LeapsAndBounds: A Method for Approximately Optimal Algorithm Configuration
* to appear at ICML 2018 
  Access Paper or Ask Questions

Linear Stochastic Approximation: Constant Step-Size and Iterate Averaging


Sep 12, 2017
Chandrashekar Lakshminarayanan, Csaba Szepesvári

Figure 1 for Linear Stochastic Approximation: Constant Step-Size and Iterate Averaging
Figure 2 for Linear Stochastic Approximation: Constant Step-Size and Iterate Averaging
* 16 pages, 2 figures, was submitted to NIPS 2017 
  Access Paper or Ask Questions

A Modular Analysis of Adaptive (Non-)Convex Optimization: Optimism, Composite Objectives, and Variational Bounds


Sep 08, 2017
Pooria Joulani, András György, Csaba Szepesvári

Figure 1 for A Modular Analysis of Adaptive (Non-)Convex Optimization: Optimism, Composite Objectives, and Variational Bounds
Figure 2 for A Modular Analysis of Adaptive (Non-)Convex Optimization: Optimism, Composite Objectives, and Variational Bounds
Figure 3 for A Modular Analysis of Adaptive (Non-)Convex Optimization: Optimism, Composite Objectives, and Variational Bounds
* Accepted to The 28th International Conference on Algorithmic Learning Theory (ALT 2017). 40 pages 
  Access Paper or Ask Questions

Mixing time estimation in reversible Markov chains from a single sample path


Aug 24, 2017
Daniel Hsu, Aryeh Kontorovich, David A. Levin, Yuval Peres, Csaba Szepesvári

* 34 pages, merges results of arXiv:1506.02903 and arXiv:1612.05330 
  Access Paper or Ask Questions

Structured Best Arm Identification with Fixed Confidence


Jun 19, 2017
Ruitong Huang, Mohammad M. Ajallooeian, Csaba Szepesvári, Martin Müller

Figure 1 for Structured Best Arm Identification with Fixed Confidence
Figure 2 for Structured Best Arm Identification with Fixed Confidence

  Access Paper or Ask Questions

Bernoulli Rank-$1$ Bandits for Click Feedback


Mar 19, 2017
Sumeet Katariya, Branislav Kveton, Csaba Szepesvári, Claire Vernade, Zheng Wen

Figure 1 for Bernoulli Rank-$1$ Bandits for Click Feedback
Figure 2 for Bernoulli Rank-$1$ Bandits for Click Feedback
Figure 3 for Bernoulli Rank-$1$ Bandits for Click Feedback

  Access Paper or Ask Questions

SDP Relaxation with Randomized Rounding for Energy Disaggregation


Oct 29, 2016
Kiarash Shaloudegi, András György, Csaba Szepesvári, Wilsun Xu

Figure 1 for SDP Relaxation with Randomized Rounding for Energy Disaggregation
Figure 2 for SDP Relaxation with Randomized Rounding for Energy Disaggregation
Figure 3 for SDP Relaxation with Randomized Rounding for Energy Disaggregation

  Access Paper or Ask Questions

(Bandit) Convex Optimization with Biased Noisy Gradient Oracles


Sep 22, 2016
Xiaowei Hu, Prashanth L. A., András György, Csaba Szepesvári

Figure 1 for (Bandit) Convex Optimization with Biased Noisy Gradient Oracles
Figure 2 for (Bandit) Convex Optimization with Biased Noisy Gradient Oracles
Figure 3 for (Bandit) Convex Optimization with Biased Noisy Gradient Oracles

  Access Paper or Ask Questions

Multiclass Classification Calibration Functions


Sep 20, 2016
Bernardo Ávila Pires, Csaba Szepesvári

Figure 1 for Multiclass Classification Calibration Functions
Figure 2 for Multiclass Classification Calibration Functions
Figure 3 for Multiclass Classification Calibration Functions
Figure 4 for Multiclass Classification Calibration Functions
* 44 pages 
  Access Paper or Ask Questions

Policy Error Bounds for Model-Based Reinforcement Learning with Factored Linear Models


Sep 20, 2016
Bernardo Ávila Pires, Csaba Szepesvári

Figure 1 for Policy Error Bounds for Model-Based Reinforcement Learning with Factored Linear Models
Figure 2 for Policy Error Bounds for Model-Based Reinforcement Learning with Factored Linear Models
* JMLR W&CP 49: COLT 2016 Proceedings (2016) 1-31 
 * 30 pages. Corrected typos. Appears in JMLR Workshop and Conference Proceedings 49: Proceedings of the 29th Annual Conference on Learning Theory (COLT 2016) 
  Access Paper or Ask Questions

Chaining Bounds for Empirical Risk Minimization


Sep 07, 2016
Gábor Balázs, András György, Csaba Szepesvári


  Access Paper or Ask Questions